To study effects of vitamin A status on retinol dynamics, male rats were fed purified diets varying in vitamin A concentration. Group 1 rats had marginal liver vitamin A levels (approximately 500 nmol) and were in a slight positive vitamin A balance; Group 2 had similar liver levels but were in a slight negative balance; Group 3 had lower liver levels (approximately 370 nmol) and were in a slight negative balance; Group 4 had depleted liver reserves (<10 nmol) and were in vitamin A balance. [3H]Retinol-labeled plasma was injected intravenously, and serial plasma samples were collected for 41 d while rats (six per group) consumed approximately 50 nmol retinol/d (Group 1) or -25 nmol/d (Groups 2-4). Plasma retinol was normal in Groups 1-3 (1.9-2.0 micromol/L) and lower in Group 4 (0.96 micromol/L). Plasma tracer data were fit to a three-compartment model. The central plasma retinol compartment (transit time, 1.5-1.7 h) exchanged with a fast turning-over extravascular vitamin A pool (transit time, 3-4.5 h; -40 nmol) and with a larger, slow turning-over extravascular pool (transit time, 5.5-10 d) that was the site of irreversible utilization of vitamin A. Irreversible utilization was 36 nmol/d (Group 1), 29 nmol/d (Groups 2 and 3) and 20 nmol/d (Group 4). The data indicate that in rats with low or marginal vitamin A status, vitamin A intake, vitamin A reserves and plasma retinol concentration all influence vitamin A utilization and other aspects of retinol dynamics.